Active/Passive Mode Enabler for Distributed Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data center systems lack integration of assets across multiple data centers, leading to operational disruptions during failures or mobility needs, especially with significant geographical distances causing delays in synchronous write IOs.
Innovation Solution
Implementing a system with logic to manage two operational modes in a distributed volume setup, allowing transitions between active/passive and active/active modes, enabling seamless mobility and data consistency across data centers without suspending operations due to link failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If synchronous write IOs are implemented across geographically dispersed data centers, then data consistency is improved, but operational delay becomes unacceptable
Solution Approach 1:
The system dynamically switches between active/active and active/passive operational modes based on real-time conditions. When data consistency is critical, it operates in active/passive mode with synchronous writes. When latency becomes problematic, it transitions to active/active mode with asynchronous writes, allowing both data centers to accept writes independently. This dynamic adaptation resolves the contradiction by adjusting the consistency model based on operational requirements.
Solution Approach 2:
The patent changes the operational parameter from synchronous to asynchronous write mode depending on the operational context. In active/passive mode, synchronous writes ensure data consistency but introduce delay. In active/active mode, asynchronous writes reduce latency but require conflict resolution mechanisms. The system selectively applies different write parameters based on the operational mode to balance consistency and performance.
2Adaptability or versatility
If assets of different data centers are integrated for normal operations, then federation and mobility are improved, but risk of operational suspension due to link failure increases
Solution Approach 1:
The system employs dynamic mode switching between active/active and active/passive configurations. In active/passive mode, the passive data center acts as a standby with minimal operational risk. In active/active mode, both data centers operate simultaneously with enhanced federation capabilities. The control logic monitors link health and automatically transitions modes to balance federation benefits against suspension risks, allowing the system to tolerate link failures better.
Solution Approach 2:
The patent implements a passive data center that serves as a pre-prepared backup before any failure occurs. This passive configuration acts as a cushion against potential link failures or data center outages. By having a standby data center ready in advance, the system can absorb shocks from link failures without suspending operations, as the passive data center can immediately take over if needed.
3Productivity
If active/active asynchronous capability is implemented, then resource utilization is improved, but data consistency complexity increases
Solution Approach 1:
The patent segments the operational modes into distinct active/active and active/passive configurations. Each segment handles data consistency differently: active/passive uses simple synchronous replication, while active/active uses asynchronous replication with conflict resolution. This segmentation allows the system to use the simpler consistency model when possible and only invokes the more complex active/active mode when resource utilization benefits outweigh the consistency management overhead.
Solution Approach 2:
The control logic acts as an intermediary that manages the complexity of data consistency in active/active mode. It coordinates between the two active data centers, resolving conflicts and ensuring eventual consistency. This intermediary layer abstracts the complexity from the individual data centers, allowing them to operate autonomously while maintaining overall system consistency through the mediating control logic.
Data Source
AI summary
Consistency groups of federated data centers can be transitioned between active/active multi-writer and active/passive single writer operational modes. In the active/passive single writer mode one or more hosts in a first data center have Write access to a distributed storage resource, but a host in a second data center does not have Write access to that distributed storage resource. In the active/active multi-writer mode hosts in both the first and the second data center have Write access to the distributed storage resource. The active/passive single writer mode is used as a default mode. The active/active multi-writer mode is used when needed for scheduled or unscheduled but qualified events which require mobility, e.g., moving virtual machines or cluster groups.